Supplementary material
Timothy J. R. Harris and Frank McCormick
Other cancers
Glioma
One of the largest DNA sequencing efforts in cancer is the Cancer Genome Atlas study sponsored by
the National Cancer Institute and National Human Genome Research Institute. The pilot project
looked at copy number, gene expression, and other molecular changes in 206 glioblastomas. An
integrated view of the glioblastoma genome from this study identified changes in genes in three major
pathways: receptor tyrosine kinase (RTK) signaling, the p53 and the retinoblastoma suppressor gene
pathways. Frequent deletions and mutations in PTEN were observed, and 86% of the samples had a
change in the RTK/PI3K pathway. Somatic mutations in the regulatory subunit PIK3RI were found in
approximately 10% of the gliomas, and it is becoming increasingly clear from other sequencing
studies that mutations in the gene coding for the catalytic subunit of PI3K (p110) are also quite
common in other cancers. Mutations in the genes coding for the two PI3K subunits seem to be
mutually exclusive. A comprehensive review of the role of PI3K in cancer cell signaling has been
published.1
Distinct mutations in the extracellular domain of EGFR predict sensitivity to EGFR tyrosine kinase
inhibitors in glioma, but only when PTEN is active.2 In the p53 pathway, there were amplifications of
MDM2 and MDM4, and frequent ARF deletions. In the retinoblastoma pathway, the most common
deletion was in the CDKN2A/CDKN2B locus on chromosome 9q21, and amplification of CDK4.
An important biomarker in glioma seems to be the methylation status of the promoter region of
MGMT, which predicts sensitivity to temozolomide, the standard of care treatment for the disease, and
has implications for treatment regimens in the future management of the disease.3
Mutations in the NF1 gene, conventionally associated with neurofibromatosis, have also been
described. Loss of NF1 is expected to lead to upregulation of the Ras/MAPK pathway, and could
sensitize these tumors to therapies based on Raf or MEK inhibitors. Another genome-wide study of
glioma showed unexpected recurrent mutations in the active site of IDH1, an enzyme involved in
glucose metabolism.4 These studies have been confirmed and extended to include the IDH2 gene,
suggesting the possibility that modulation of IDH activity may provide some therapeutic opportunity
for those patients in whom these enzymes are inactivated.5 This result emphasizes the advantages of
technical rather than hypothesis-driven research as IDH1 was never considered to be a target for
intervention until mutations in this gene were found by sequencing. The mutations are in fact gain of
function mutations, which change the substrate specificity of the IDH 1 enzyme.6
Hepatocellular carcinoma
Cancer of the liver is associated with environmental insults, such as toxin exposure and chronic
hepatitis B and C virus infection. This type of cancer is more common in Africa and Asia than in
Europe and the US. One of the problems of this disease is predicting recurrence after treatment, which
results in cure in some patients, but not in others. Formalin-fixed, paraffin-embedded tissue (FFPET)
is the predominant source of archived liver tumor samples and is available for most patients at the time
clinical outcome becomes available. Gene expression in FFPET from hepatocellular carcinoma
biopsies demonstrated that samples from 90% of the patients yielded data of high quality despite some
of the samples being over 20 years old. The study supports that FFPET can be used to derive
transcriptional profiles that correlate expression signature with outcome.7
Hepatocellular carcinoma cells frequently contain genetic abnormalities, including chromosomal
deletions, amplifications and loss of heterozygosity, some of which involve known oncogenes. In a
comprehensive study of the molecular aberrations in hepatocellular carcinoma, overexpression of
VEGF via 6p21 gain was identified.8 Disruption of WNT-β catenin signaling via mutations in AXIN1
1
or CTNNB1 is also involved in the development of hepatocellular carcinoma.9 Dividing patients into
subtypes using gene expression levels may be possible, since 50% of tumors have WNT or AKT
pathway activation.10 Sorafenib, an inhibitor of several tyrosine kinases that targets several proteins in
different signaling pathways, including Raf1, B-Raf, VEGFR2, PDGFR, and c-kit, has shown efficacy
in the treatment of this disease.11 Aberrant activation of the Ras/Raf/ERK pathway in hepatocellular
carcinoma at least in part occurs through decreased levels of the Sprouty-related suppressor of Ras
protein Spred, and expression of this protein is inversely correlated with the incidence of tumor
invasion and metastasis.12 This may account, in part, for the sensitivity of hepatocellular carcinoma to
sorafenib, since patients with elevated MAPK activity show better clinical responses to this agent.
Despite the fact that combination therapies will still be needed, the prospect of being able to categorize
patients with hepatocellular carcinoma based on the activation status of different signaling pathways in
the tumor cells is an exciting therapeutic prospect.13
Multiple myeloma
Multiple myeloma is a malignancy of plasma cells with a variable outcome following standard or
high-dose treatment. Myelomas are usually categorized by chromosome analysis, and can be
segregated into nonhyperploid or hyperploid types. In the latter, there are often trisomies in
chromosomes 3, 5, 11,15, 19, and 21, which can be subdivided by microarray analysis into tumors that
have the trisomies above (except 11), and have additional chromosomal alterations, including gains of
chromosome 1q and chromosome 7, and deletion of chromosome 13. Nonhyperploid disease can be
subdivided into two groups: one with deletions of chromosomes 8 and 13 and one with amplification
of chromosome 1q and deletion of 1p and 13. Translocation in chromosome 14q32 involving several
genes is also apparent in up to 40% of myeloma patients.
A group at the Myeloma Institute in Arkansas, USA has extended the classification of myeloma
using gene-expression arrays.14 In their study, two high-risk profiles were obtained based on
chromosome 14q32 translocations and the expression of genes during proliferation. This builds on
data pointing to the importance of chromosome 1 abnormalities, including the fact that gains of 1q21
lead to inferior survival. This approach may provide a more powerful and cheaper alternative to the
extensive chromosomal analysis used to identify those at higher risk for early fatal disease.
Gene-expression profiling in myeloma has become an important component of the clinical
monitoring of patients treated with bortezomib, which is an inhibitor of NF-κB activation. The geneexpression stratification for patients with newly diagnosed multiple myeloma treated with high-dose
chemotherapy is predictive of outcome in those with relapsed disease treated with bortezomib or highdose steroids.15,16 Gene-expression is also being used to assess patients with respect to proliferationfree survival time after relapse.17 More information on the somatic mutations found in myeloma cells
will be forthcoming from the whole genome sequencing studies sponsored by the Multiple Myeloma
Research Consortium. These studies will point to other drugs that can be used in subsets of patients,
based on the genetic lesions in their tumor cells.
Ovarian cancer
Ovarian cancer cure rates are low, owing to the generally late presentation of patients with the disease.
In cases that are detected early, treatment with taxol and platinum-based compounds leads to a good
outcome. Some screening, using transvaginal sonography and monitoring of CA125 levels is available
for high-risk patients. Other proteomic tests adding sensitivity to simple CA125 testing are in
development (see http://www.vermillion.com). Ovarian cancer can be differentiated into low-grade
and high-grade tumors based on genetic information. High-grade tumors are often associated with
BRCA1 and BRCA2 mutations and loss of heterozygosity on chromosomes 7q and 9p. Mutation and
loss of TP53 occur in a high proportion of both familial and sporadic tumors. Low-grade tumors have
mutations in KRAS, BRAF and PI3KC.18 Analysis of p53 and KRAS to differentiate low-grade serous
and non-serous type I cancers from high-grade serous and endometroid type II cancer, would be useful
from a treatment perspective.18 Type I cancers should respond better to drugs targeting the PI3K and
Ras-MAPK pathways than type II cancer.19 Nevertheless, the clinical management of ovarian cancer
targeting mutations in the PI3K pathway will be complex. It will be important to look for mutations in
PI3K and the regulatory subunit, and also PTEN deletions, AKT mutations, and amplifications of other
genes, such as AKT2. Dual PI3K-mTOR inhibitors, isoform specific PI3K inhibitors, and specific
2
AKT or mTOR-targeted molecules may have potential use. Several compounds of this kind are in
clinical development.1,20
Low levels of two different microRNAs (Dicer and Drosha), involved in splicing pre-mRNA, have
been associated with advanced ovarian tumors and are known markers of poor prognosis.21
In an uncommon form of ovarian cancer (granulosa cell tumors), transcriptome sequencing has
shown that mutation of the transcription factor FOXL2, is a potential driver for the uncontrolled
proliferation of granulosa cells. Analysis of this gene for mutations might be useful to aid the
diagnosis of this particular form of ovarian cancer.22 Although the Cancer Genome Atlas consortium
results have not yet been published, accumulating data suggest that there are several molecular
subtypes of ovarian cancer, based on loss of suppressor genes or patterns of mutations in genes in
different signaling pathways (http://cancergenome.nih.gov).
Pancreatic cancer
Pancreatic cancer is one of the most fatal cancers, owing primarily to lack of early detection. Several
genetic lesions have been found in these cancers, including mutations in KRAS, TP53, CDKN2A, and
SMAD4. Genome-wide studies of gene expression and genome organization in 24 pancreatic tumors
have revealed many point mutations, which defined alterations in a core set of 12 signaling
pathways.23 The pathways included: apoptosis, DNA repair, cell-cycle control, Wnt/Notch pathway,
Hedgehog signaling, and KRAS. The implications for therapy are important, given that the studies
highlight the Hedgehog and the Wnt/Notch pathways, both of which are targets for new therapies. 24
Candidate gene studies have implicated PALB2 in familial pancreatic cancer.25 The other major
familial pancreatic cancer susceptibility gene is BRCA2 (the product of which binds to the PALB2
protein), implicating DNA repair abnormalities in this disease. Pancreatic cancer patients with
abnormal signaling through these pathways might, therefore, benefit from exposure to therapies
targeting DNA repair mechanisms such as PARP inhibitors.26
Renal cell carcinoma
Studies of familial renal cell carcinoma (RCC) have allowed a basic understanding of the molecular
basis of the sporadic disease, and led to more rational treatment regimens. 27 Four major familial forms
of the disease exist, each one of which has contributed useful and connected information. Von HippelLindau disease patients suffer from clear cell RCC that is primarily caused by dominant germline
mutations in the VHL tumor suppressor gene on chromosome 3. Inactivation of the suppressor has
multiple effects, including interference with the degradation of HIF1A. Papillary renal cell carcinoma
is the second most common form of the disease and can be subdivided into two types, based on geneexpression profile and histopathology. Mutations in MET, the receptor for hepatocyte growth factor
(HGF), are common in type I hereditary papillary renal cell carcinoma, and were discovered through
classic linkage genetics.
Since that time, activation of c-Met signaling through mutation and overexpression has been
described not only in hereditary renal cancer but also in sporadic forms of the disease and in many
other cancers. Many different mutations in the tyrosine kinase domain of the c-Met protein have been
described. This observation indicates that the MET gene should be sequenced in all RCC patients to
determine mutation and/or amplification status.
The HGF/c-Met interaction has suggested several ways forward clinically for papillary renal cell
carcinoma, including the development of anti-HGF monoclonal antibodies. Given that several
approved tyrosine kinase inhibitors exist, the development of combination therapies targeting the
particular pathways activated in RCC, including not only c-Met but also enzymes downstream of
c-Met, is a real clinical option. Sunitinib, an effective inhibitor of c-Met, is now being used as firstline therapy to treat RCC, pazopanib from GlaxoSmithKline (Brentford, UK), a more selective
tyrosine kinase inhibitor, has received a favorable review from the Oncologic Drugs Advisory
Committee for the treatment of RCC. XL880, a multikinase and a potent c-Met inhibitor from Exelixis
(San Francisco, CA, USA), is being assessed in phase II clinical trials in RCC. As in other
malignancies, the potential to combine these drugs with rapamycin analogs that target mTOR is
particularly encouraging.
One of the other forms of hereditary RCC is caused by inactivation of fumarate dehydratase
(hereditary leiomyomatosis and RCC), which may suggest that further study of genes coding for
3
metabolic enzymes in cancer is warranted, given the findings of IDH1 mutations in both glioma and
acute myeloid leukemia.28
Supplementary reference list
1. Engelman, J. A. Targeting PI3K signaling in cancer: opportunities, challenges and limitations. Nat.
Rev. Cancer 9, 550–562 (2009).
2. Mellinghoff, I. K. et al. Molecular determinants of the response of glioblastomas to EGFR kinase
inhibitors. N. Engl. J. Med. 353, 2012–2024 (2005).
3. Brandes, A. A. et al. Correlations between 06-methylguanine DNA methyltransferase promoter
methylation status, 1p and 19q deletions, and response to temozolomide in anaplastic and recurrent
oligodendroglioma: a prospective GICNO study. J. Clin. Oncol. 24, 4746–4753 (2006).
4. Parsons, D. W. et al. An integrated genomic analysis of human glioblastoma multiforme. Science
321, 1807–1812 (2008).
5. Yan, H. et al. IDH1 and IDH2 mutations in gliomas. N. Engl. J. Med. 360, 765–773 (2009).
6. Dang, L. et al. Cancer associated IDH1 mutations produce 2-hydroxyglutarate. Nature 462, 739–
744 (2009).
7. Hoshida, Y. et al. Gene expression in fixed tissues and outcome in hepatocellular carcinoma. N.
Engl. J. Med. 359, 1995–2004 (2008).
8. Chiang, D. Y. et al. Focal gains of VEGFA and molecular classification of hepatocellular carcinoma.
Cancer Res. 68, 6779–6788 (2008).9. Mínguez, B., Tovar, V., Chiang, D., Villaneva, A. & Llovet,
J. M. Pathogenesis of hepatocellular carcinoma and molecular therapies. Curr. Opin.
Gastroenterol. 25, 186–194 (2009).
10. Boyault, S. et al. Transcriptome classification of HCC is related to gene alterations and to new
therapeutic targets. Hepatology 45, 42–52 (2007).
11. Llovet, J. M. et al. Sorafenib in advanced hepatocellular carcinoma. N. Engl. J. Med. 359, 378–390
(2008).
12. Yoshida, T. et al. Spreds, inhibitors of the RAS/ERK signal transduction, are dysregulated in human
hepatocellular carcinoma and linked to the malignant phenotype of tumors. Oncogene 25, 6056–
6066 (2006).
13. Llovet, J. M. & Bruix, J. Molecularly targeted therapies in hepatocellular carcinoma. Hepatology 48,
1312–1327 (2008).
14. Shaughnessy, J. D., Jr. et al. A validated gene expression model of high-risk multiple myeloma is
defined by deregulated expression of genes mapping to chromosome 1. Blood 109, 2276–2284
(2007).
15. Zhan, F., Barlogie, B., Mulligan, G., Shaughnessy, J. D., Jr. & Bryant, B. High-risk myeloma: a gene
expression based risk-stratification model for newly diagnosed multiple myeloma treated with highdose therapy is predictive of outcome in relapsed disease treated with single-agent bortezomib or
high-dose dexamethasone. Blood 111, 968–969 (2008).
16. Mulligan, G. et al. Gene expression profiling and correlation with outcome in clinical trials of the
proteasome inhibitor bortezomib. Blood 109, 3177–3188 (2009).
17. Nair, B. et al. Gene expression profiling of plasma cells at myeloma relapse from tandem
transplantation trial Total Therapy 2 predicts subsequent survival. Blood 113, 6572–6575 (2009).
18. Bast, R. C., Jr., Hennessey, B. & Mills, G. B. The biology of ovarian cancer: new opportunities for
translation. Nat. Rev. Cancer 9, 415–428 (2009).
19. Crijns, A. P. et al. Survival-related profile, pathways, and transcription factors in ovarian cancer.
PLoS Med. 6, e24 (2009).
20. Liu, P. et al. Targeting the phosphoinositide 3-kinase pathway in cancer. Nat. Rev. Drug Discov. 8,
627–644 (2009).
21. Merritt, W. M. et al. Dicer, Drosha, and outcomes in patients with ovarian cancer. N. Engl. J. Med.
359, 2641–2650 (2008).
22. Shah, S. P. et al. Mutation of FOXL2 in granulosa-cell tumours of the ovary. N. Engl. J. Med. 360,
2719–2729 (2009).
23. Jones, S. et al. Core signaling pathways in human pancreatic cancers revealed by global genomic
analyses. Science 321, 1801–1806 (2008).
24. Scales, S. J. & de Sauvage, F. J. Mechanisms of Hedgehog pathway activation in cancer and
implications for therapy. Trends Pharmacol. Sci. 30, 303–312 (2009).
25. Jones, S. et al. Exomic sequencing identifies PALB2 as a pancreatic cancer susceptibility gene.
Science 324, 217 (2009).26. Xia, B. et al. Control of BRCA2 cellular and clinical functions by a
nuclear partner, PALB2. Mol. Cell 22, 719–729 (2006).
4
27. Guibellino, A., Linehan, W. M. & Bottaro, D. P. Targeting the Met signaling pathway in renal cancer.
Expert Rev. Anticancer Ther. 9, 785–793 (2009).
28. Downing, J. R. Cancer genomes—continuing progress. N. Engl. J. Med. 361, 1111–1112 (2009).
5